[William Sommerwerck]

OK, I'll bite, what is the *exact* time constant for lowest
ripple, since you claim that is the important spec.

Infinity, of course.

That would be impossible, because A: the capacitor would never charge up,
and B: the load sets the upper limit of the R in RC.

[Don Pearce[_3_]]

On Thu, 30 Aug 2012 16:28:09 +1000, "Trevor" wrote:


"Scott Dorsey" wrote in message
...
In article , Trevor
wrote:
"Don Pearce" wrote in message
...

The time constant is precisely the salient issue in power supply
design. It is the time constant that determines the amplitude of the
ripple.

OK, I'll bite, what is the *exact* time constant for lowest ripple, since
you claim that is the important spec.

Infinity, of course.


That would imply a large source resistance (as well as a large capacitance)
NOT something you want in a power supply design.
Try again ;-)

Trevor.

What do you mean by a large source resistance? A
transformer/diode/capacitor power supplier is not linear, it is a
switcher. When the topping-up phase at the tip of the mains wave is
done, the diodes switch off, leaving the source resistance infinite.

d

[William Sommerwerck]

What do you mean by a large source resistance? A
transformer/diode/capacitor power supplier is not linear,
it is a switcher.

Ouch!

When the topping-up phase at the tip of the mains wave
is done, the diodes switch off, leaving the source resistance
infinite.

Not as far as the load is concerned.

[Don Pearce[_3_]]

On Thu, 30 Aug 2012 09:49:31 -0700, "William Sommerwerck"
wrote:

What do you mean by a large source resistance? A
transformer/diode/capacitor power supplier is not linear,
it is a switcher.

Ouch!

When the topping-up phase at the tip of the mains wave
is done, the diodes switch off, leaving the source resistance
infinite.

Not as far as the load is concerned.


The load sees the capacitor. The capacitor sees a switcher.

d

[William Sommerwerck]

When the topping-up phase at the tip of the mains
wave is done, the diodes switch off, leaving the
source resistance infinite.

Not as far as the load is concerned.

The load sees the capacitor. The capacitor sees a switcher.

The capacitor is able to supply current to the load. It therefore does not
present an infinite source impedance.

[Don Pearce[_3_]]

On Thu, 30 Aug 2012 11:48:24 -0700, "William Sommerwerck"
wrote:

When the topping-up phase at the tip of the mains
wave is done, the diodes switch off, leaving the
source resistance infinite.

Not as far as the load is concerned.

The load sees the capacitor. The capacitor sees a switcher.

The capacitor is able to supply current to the load. It therefore does not
present an infinite source impedance.


No the capacitor sees an infinite source impedance, it doesn't present
one. During most of the operational cycle the amplifier itself is the
only impedance seen by the capacitor.

d

[Trevor]

"Don Pearce" wrote in message
...
On Thu, 30 Aug 2012 16:28:09 +1000, "Trevor" wrote:
"Scott Dorsey" wrote in message
...
In article , Trevor
wrote:
"Don Pearce" wrote in message
...

The time constant is precisely the salient issue in power supply
design. It is the time constant that determines the amplitude of the
ripple.

OK, I'll bite, what is the *exact* time constant for lowest ripple,
since
you claim that is the important spec.

Infinity, of course.


That would imply a large source resistance (as well as a large
capacitance)
NOT something you want in a power supply design.
Try again ;-)

What do you mean by a large source resistance? A
transformer/diode/capacitor power supplier is not linear, it is a
switcher.

Leaves me to wonder what value you use to calculate the Time Constant you
claim to start with then? How do you calculate the value for C if you use
infinite R! :-)
And saying the R is non linear kind of blows your whole start with the time
constant design idea out the window doesn't it?


When the topping-up phase at the tip of the mains wave is
done, the diodes switch off, leaving the source resistance infinite.

Surely the time constant only applies during charge, "top up phase", so you
can't use an infinite R to calculate the time constant.

Trevor.

[Don Pearce[_3_]]

On Fri, 31 Aug 2012 08:40:06 +1000, "Trevor" wrote:


"Don Pearce" wrote in message
...
On Thu, 30 Aug 2012 16:28:09 +1000, "Trevor" wrote:
"Scott Dorsey" wrote in message
...
In article , Trevor
wrote:
"Don Pearce" wrote in message
...

The time constant is precisely the salient issue in power supply
design. It is the time constant that determines the amplitude of the
ripple.

OK, I'll bite, what is the *exact* time constant for lowest ripple,
since
you claim that is the important spec.

Infinity, of course.


That would imply a large source resistance (as well as a large
capacitance)
NOT something you want in a power supply design.
Try again ;-)

What do you mean by a large source resistance? A
transformer/diode/capacitor power supplier is not linear, it is a
switcher.

Leaves me to wonder what value you use to calculate the Time Constant you
claim to start with then? How do you calculate the value for C if you use
infinite R! :-)
And saying the R is non linear kind of blows your whole start with the time
constant design idea out the window doesn't it?


When the topping-up phase at the tip of the mains wave is
done, the diodes switch off, leaving the source resistance infinite.

Surely the time constant only applies during charge, "top up phase", so you
can't use an infinite R to calculate the time constant.

Trevor.


There are two time constants, both involving the big reservoir cap.
The first is the charging time constant, which is the sum of the
resistances of the transformer winding and the diodes. The second is
the discharge constant which the current drain of the amplifier
itself. Both of these are finite.

How exactly do you design power supplies? It sounds like you are
clueless how they work, and you just guess.

d

[Trevor]

"Don Pearce" wrote in message
...
The time constant is precisely the salient issue in power supply
design. It is the time constant that determines the amplitude of the
ripple.

OK, I'll bite, what is the *exact* time constant for lowest ripple,
since
you claim that is the important spec.

Infinity, of course.


That would imply a large source resistance (as well as a large
capacitance)
NOT something you want in a power supply design.
Try again ;-)

What do you mean by a large source resistance? A
transformer/diode/capacitor power supplier is not linear, it is a
switcher.

Leaves me to wonder what value you use to calculate the Time Constant you
claim to start with then? How do you calculate the value for C if you use
infinite R! :-)
And saying the R is non linear kind of blows your whole start with the
time
constant design idea out the window doesn't it?


When the topping-up phase at the tip of the mains wave is
done, the diodes switch off, leaving the source resistance infinite.

Surely the time constant only applies during charge, "top up phase", so
you
can't use an infinite R to calculate the time constant.


There are two time constants, both involving the big reservoir cap.
The first is the charging time constant, which is the sum of the
resistances of the transformer winding and the diodes. The second is
the discharge constant which the current drain of the amplifier
itself. Both of these are finite.

Right, so what? That was never disputed, just why you would start with
either time constant.


How exactly do you design power supplies? It sounds like you are
clueless how they work, and you just guess.


Well at least I don't claim I start with the time constant and then be
unable to say what that figure is or how you calculate the required cap size
from it. Obviously you have never tried to do what you claim and are just
continuing to argue for the sake of it. What you gain from that is anybody's
guess, but I'm done.

Trevor.

[Don Pearce[_3_]]

On Fri, 31 Aug 2012 20:29:14 +1000, "Trevor" wrote:


"Don Pearce" wrote in message
...
The time constant is precisely the salient issue in power supply
design. It is the time constant that determines the amplitude of the
ripple.

OK, I'll bite, what is the *exact* time constant for lowest ripple,
since
you claim that is the important spec.

Infinity, of course.


That would imply a large source resistance (as well as a large
capacitance)
NOT something you want in a power supply design.
Try again ;-)

What do you mean by a large source resistance? A
transformer/diode/capacitor power supplier is not linear, it is a
switcher.

Leaves me to wonder what value you use to calculate the Time Constant you
claim to start with then? How do you calculate the value for C if you use
infinite R! :-)
And saying the R is non linear kind of blows your whole start with the
time
constant design idea out the window doesn't it?


When the topping-up phase at the tip of the mains wave is
done, the diodes switch off, leaving the source resistance infinite.

Surely the time constant only applies during charge, "top up phase", so
you
can't use an infinite R to calculate the time constant.


There are two time constants, both involving the big reservoir cap.
The first is the charging time constant, which is the sum of the
resistances of the transformer winding and the diodes. The second is
the discharge constant which the current drain of the amplifier
itself. Both of these are finite.

Right, so what? That was never disputed, just why you would start with
either time constant.


How exactly do you design power supplies? It sounds like you are
clueless how they work, and you just guess.


Well at least I don't claim I start with the time constant and then be
unable to say what that figure is or how you calculate the required cap size
from it. Obviously you have never tried to do what you claim and are just
continuing to argue for the sake of it. What you gain from that is anybody's
guess, but I'm done.

Trevor.


Last time try. The time constant determines how far the voltage sags
between recharges. It is as simple as that. You have a ripple spec -
you choose a time constant that meets it. How you have failed to grasp
this beats me.

d